Survivability Strategies for Emerging Wireless Networks With Data Mining Techniques: a Case Study With NetLogo and RapidMiner

Abstract

Emerging wireless networks have brought Internet and communications to more users and areas. Some of the most relevant emerging wireless technologies are Worldwide Interoperability for Microwave Access, Long-Term Evolution Advanced, and ad hoc and mesh networks. An open challenge is to ensure the reliability and robustness of these networks when individual components fail. The survivability and performance of these networks can be especially relevant when emergencies arise in rural areas, for example supporting communications during a medical emergency. This can be done by anticipating failures and finding alternative solutions. This paper proposes using big data analytics techniques, such as decision trees for detecting nodes that are likely to fail, and so avoid them when routing traffic. This can improve the survivability and performance of networks. The current approach is illustrated with an agent-based simulator of wireless networks developed with NetLogo and data mining processes designed with RapidMiner. According to the simulated experimentation, the current approach reduced the communication failures by 51.6% when incorporating rule induction for predicting the most reliable routes.